Meta’s Muse AI tool is gone: What it taught us about Instagram privacy

If you use Instagram, you may have heard about an AI feature Meta launched and pulled within days. It was called Muse — an image generation tool that could produce new photos based on your existing ones. The backlash was swift, and Meta removed the feature on July 10, 2026, admitting it “missed the mark.” But the episode raised real questions about how your photos could be used for AI training and what control you actually have.

Here’s what happened, why it mattered, and what you can do to protect your images going forward.

What happened

Muse was designed to let Instagram users create new images by feeding the AI a photo of themselves or others. In theory, you could generate stylized portraits or backgrounds that looked like you. But the controversy centered on where the AI got its source material. Meta’s terms allowed it to use public Instagram photos to train its generative AI models, including Muse. That meant any photo you posted publicly could be used — without your explicit permission — to create new images that resembled you.

The tool was announced in late June 2026 and rolled out in early July. Almost immediately, privacy advocates and user groups raised concerns about consent and likeness rights. On July 9, SAG-AFTRA, the actors’ union, recommended that its members opt out of Meta’s AI features entirely, warning that the tool could replicate performers’ likenesses without compensation. The union’s statement said, “Take action to protect your likeness.”

By July 10, Meta had removed the feature. In a statement reported by Yahoo Tech, the company said Muse “missed the mark” and that it would take more time to address user feedback.

Why it matters for Instagram users

The Muse episode is not an isolated incident. It’s part of a broader push by Meta to integrate generative AI into its platforms. Even though Muse is gone, the underlying data practices still apply. Meta’s privacy policy says it may use your photos and other content to train AI models unless you opt out. That opt-out process has been criticized for being hard to find and not clearly explained.

The core issue is whether you want your appearance — your face, your body, your style — used to train a system that could later generate images without your direct involvement. Unlike a private message, a public Instagram photo is fair game under Meta’s current terms. And once your likeness is in the training data, there is no way to remove it retroactively.

Also, this may not be the last time Meta tries something like Muse. The company has invested heavily in AI advertising tools and image generation for business accounts. The removal of Muse may be temporary, or it may lead to a version with clearer consent. But for now, the privacy risk remains.

What you can do to protect your photos

You have some control, though the options are limited and not always straightforward. Here are concrete steps you can take right now.

Opt out of Meta’s AI training. On Instagram, go to your account settings, then Privacy, and look for a section called “Data use” or “AI training.” (The exact label may change over time.) Meta allows you to object to its use of your content for generative AI. You may need to submit a request and wait for confirmation. This opt-out should apply to future training, but it cannot undo data already used.

Make your account private. If your Instagram account is set to public, any photo or video you post can be used for AI training under Meta’s policy. Switching to a private account limits what Meta can legally scrape, though it may still use data from accounts you interact with. Go to Settings > Privacy > Account Privacy and toggle on “Private account.”

Review your past posts. Even with a private account, old public posts remain visible. Meta can still access content that was public at the time you posted it. Consider deleting or archiving older photos you do not want used for AI purposes. You can bulk delete or archive from the “Your Activity” section.

Watermark or lower resolution. There is no guarantee this stops AI training, but posting images with visible watermarks or at lower resolution can make them less useful for training high-quality models. This is a stopgap, not a solution.

Limit metadata. Remove location data and other embedded information from your photos before uploading. While this does not affect the image itself, it reduces the data tied to your content.

Stay informed. Meta’s policies change frequently. Bookmark the company’s AI privacy page and check for updates before using new features. If you use Instagram for business or as a creator, consider consulting a digital rights organization like the Electronic Frontier Foundation for the latest guidance.

Sources

  • Yahoo Finance: “What Meta’s Muse AI image tool means for Instagram privacy” (July 10, 2026)
  • Yahoo Tech: “Meta Removes Muse Image AI Feature After Backlash: ‘Missed The Mark’” (July 10, 2026)
  • Yahoo: “SAG-AFTRA Recommends Members Opt-Out Of Meta’s AI Feature” (July 9, 2026)
  • MediaPost: “Meta To Integrate ‘Reasoning’ Image-Generation Tool Into AI Ad Suite” (July 8, 2026)